The xFoundry Alliance is transforming universities into solution engines that generate real ventures, develop entrepreneurial leaders, and advance meaningful societal outcomes. Headquartered at the University of Maryland's E.A. Fernandez IDEA Factory, xFoundry empowers students to tackle humanity's most pressing challenges through multidisciplinary collaboration, AI-powered tools, and rigorous venture-building programs.
Our flagship Xperience program is a 15â18-month venture-building competition in which student teams develop investment-grade MVPs that address grand societal challenges in climate, education, food and water, health, infrastructure, and society. Winning teams compete for investments of up to $2 million and receive guidance from seasoned CXOs through our Xcelerate accelerator program.
We are seeking an Associate Director of Development & AI Systems Architect to lead the design, development, and deployment of the intelligent infrastructure that powers xFoundryâs platforms. This is a hands-on technical leadership role for someone who wants to architect the AI systems that support university-driven entrepreneurshipâ from multi-agent orchestration and knowledge retrieval to enterprise-grade deployment pipelines that operate in both cloud and airgapped environments.
You will own the end-to-end AI architecture: designing how models are routed and composed, how context is managed and compressed across long-running conversations, how agents coordinate in parallel to accomplish complex tasks, and how organizational knowledge is ingested, indexed, and retrieved. This is not a research-only roleâ you will build and ship production systems that students, mentors, and partners rely on every day.
This role is ideal for someone who:
- Thinks in systems. You see how model routing, context windows, retrieval pipelines, and agent orchestration fit together as a unified architecture
- Has built and deployed LLM-powered applications in production, not just prototyped them
- Understands the tradeoffs between vector search, graph-based retrieval, and hybrid approaches, and knows when each is appropriate
- Gets energized by the challenge of making AI infrastructure work reliably in constrained environments, including enterprise and airgapped deployments
- Can design agent architectures that coordinate parallel sub-agents, manage shared state, and handle long-running tasks gracefully
- Cares about building tools that genuinely empower students and educators, not just impressive demos
AI Infrastructure & Architecture
- Design and maintain the core AI platform architecture, including model routing and provider abstraction layers that allow seamless switching between commercial and open-source models based on task requirements, cost, and deployment constraints
- Architect context management and compaction systems that enable productive, long-running interactions, ensuring conversations retain relevant history while operating efficiently within model context limits
- Build and optimize retrieval-augmented generation pipelines encompassing document and datasource ingestion, chunking strategies, embedding generation, and retrieval across both vector databases and knowledge graphs
- Define infrastructure patterns for enterprise and airgapped deployments, ensuring xFoundryâs AI capabilities can operate in environments with restricted network access and elevated security requirements
Agent Orchestration & Execution
- Design multi-agent orchestration systems where specialized agents coordinate to accomplish complex tasks, including parallel sub-agent execution, inter-agent communication, and state management across long-running workflows
- Implement skill and tool use frameworks that give agents the ability to discover, invoke, and compose external tools and services through well-defined interfaces and policy-governed registries
- Architect sandboxed execution environments for agents operating in cloud and hosted settings, ensuring secure code execution, resource isolation, and proper lifecycle management
- Build scheduled and asynchronous task execution systems that allow agents to perform recurring work, process background jobs, and manage long-running operations without blocking user interactions
Knowledge & Data Systems
- Design and manage document and datasource ingestion pipelines that process diverse source material, including PDFs, web pages, audio, and structured data, into indexed, retrievable knowledge stores
- Architect knowledge graph construction and retrieval systems that capture entity relationships, enable graph-based traversal, and support semantic queries across organizational knowledge
- Build shared filesystem management for organizations, implementing permission-based access controls, organizational scoping, and secure storage patterns that support multi-tenant AI workloads
- Develop integration architectures that connect AI systems with external services through encrypted credential management, connection health monitoring, and extensible connector frameworks
Platform & Team Leadership
- Collaborate with the broader development team to ensure AI systems integrate cleanly with frontend interfaces, backend services, and data infrastructure
- Establish engineering standards for AI system development, including testing strategies, deployment practices, monitoring, and documentation
- Evaluate emerging technologies and architectural patterns, making pragmatic recommendations about when to adopt, adapt, or build from scratch
- Contribute to product strategy by translating technical capabilities and constraints into clear guidance for stakeholders, helping shape whatâs possible and whatâs practical
â
Production AI Systems Experience
You have designed and deployed LLM-based systems that real users depend on. You understand the full lifecycle, from prompt engineering and model selection through context management, retrieval integration, and production monitoring. You know the difference between a working prototype and a reliable system.
Deep Retrieval & Knowledge Architecture Expertise
You can architect end-to-end knowledge pipelines: ingesting diverse documents, choosing appropriate chunking and embedding strategies, building retrieval systems that combine vector search with graph-based approaches, and tuning retrieval quality for real-world use cases. You understand when dense retrieval, sparse retrieval, or hybrid approaches are appropriate.
Agent Orchestration & Systems Design
You think naturally in terms of multi-agent systems: how to decompose complex tasks, coordinate parallel execution, manage shared context and state, and handle failure gracefully. You can design agent architectures that are robust, observable, and maintainable.
â
Preferred Qualifications
- Experience with enterprise or airgapped deployment environments and the unique constraints they impose
- Background in knowledge graph construction, entity resolution, and graph-based retrieval
- Familiarity with streaming architectures for real-time AI interactions (e.g., server-sent events, WebSocket-based systems)
- Experience building developer tools, SDKs, or platform infrastructure for AI systems
- Contributions to open-source AI/ML projects or published research in relevant areas
- Experience mentoring engineers or leading technical teams in fast-paced environments
- Understanding of security best practices for AI systems, including credential management and sandboxed execution
â
Build the Infrastructure That Powers Innovation: Architect the AI systems that enable the next generation of student-led ventures tackling climate, health, education, and societyâs grand challenges
Real Impact, Not Research Theater: Your systems will be used daily by students, mentors, and partnersâthis is production engineering with a mission
Technical Depth With Breadth: Work across the full AI stack, from model orchestration and retrieval systems to agent execution and enterprise deployment
Shape the Architecture: Join at a stage where your architectural decisions will define how the platform scalesâthis isnât maintenance, itâs creation
Competitive Compensation: Â $155,000 â $175,000 salary with benefits
Submit your resume along with a brief statement (1 page max) addressing the following:
mtech-human-resources@umd.edu
- Describe an AI system youâve built that involved retrieval-augmented generation or multi-agent orchestration. What architectural decisions did you make, what tradeoffs did you navigate, and what would you do differently today?
- A technical challenge in deploying LLM-based systems to production that you found particularly interesting to solve, and your approach
- What excites you about building AI infrastructure that supports university-driven entrepreneurship and venture development?
The University of Maryland is an equal opportunity employer. All qualified applicants will receive consideration for employment without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, disability, or protected veteran status.
The University of Maryland will not sponsor the successful candidate for work authorization in the United States now or in the future. UMD will not support F-1 STEM OPT for this position.
College Park, MD or Dallas, TX